Show simple item record

dc.contributor.advisorArnault, Malcolm
dc.contributor.authorHolmgren, Gary L.en_US
dc.date.accessioned2019-10-11T15:11:26Z
dc.date.available2019-10-11T15:11:26Z
dc.date.created1964en_US
dc.date.issued1964en_US
dc.identifieraleph-254745en_US
dc.identifier.urihttps://repository.tcu.edu/handle/116099117/34622
dc.description.abstractThe present research, preceded by the speaker recognition experiments of McGehee (1937), Pollack, Pickett, and Sumby (1954), Peters (1954, 1956), Howell (1960), Meeker and Nelson (1962 ), and McGee (1961), was concerned with two major questions: (1) How can the quality or judged voice characteristics of speakers' voices be measured and described?; (2) In what way do judged voice characteristics relate to actual physical voice characteristics ? The intent of the present research was to look, not specifically at the ability of listeners to identify speakers by name, but rather at a more fundamental question, how do listeners differentiate among speakers? The assumption here, which relates to the task of speaker recognition, is that when listeners identify a particular speaker by voice, they do so on the basis of responding to a particular set of cues to which they attend when hearing the speaker's voice. Likewise, when listeners differentiate among speakers, they do so on the basis of the extent to which the cues present in one speaker's voice differ from the cues present in another. It was felt, therefore, that before the actual task of identifying speakers by name can be investigated, one should initially study the extent to which listeners can consistently differentiate among speakers and in the process identify, if possible, the differential characteristics listeners attribute to the speakers' voices and then relate the judged voice characteristics to the speaker's actual physical characteristics. Utilizing Osgood's semantic differential rating form technique (Osgood, Suci, and Tannenbaum, 1957) two experiments were conducted in order to arrive at a set of reliable listener judgments of various speakers' voices on certain voice characteristics. Physical voice characteristic measures supplied by Air Force Cambridge Research Laboratories (AFCRL) included mean and variance of voiced amplitude, mean and variance of unvoiced amplitude; mean and variance of pitch, and a duration measure. The hypotheses predicted that the rating form technique would yield reliable judged voice characteristics and that these would be correlated with various physical voice characteristics. In each of the two experiments, the rating form developed (Semantic Differential Rating Form IV) was employed by 10 listeners to rate 10 speakers from AFCRL's speaker library. The experimental design provided for replication over groups of listeners, for replication over three days of testing and for counter-balancing the order of speaker presentation. A familiarization rating task was employed before each of the voice characteristic rating tasks in the first experiment but not in the second. The results of the present research revealed that listeners can reliably differentiate among speakers' voices in terms of judged voice characteristics. Judged voice characteristics can be accounted for by only two basic factors. Factor I was based on a composite of the following descriptive terms: complex. fast, high, shallow, sharp. whining, hard, busy, and intense. Factor II was a composite of ugly, dirty, and bad. Factor I was found to be significantly correlated with the composite physical measures of mean voiced amplitude and mean pitch (R mult. = . 8073, P < 05). Factor II was found to be significantly correlated with the composite physical measures of mean voiced amplitude and variance of voiced amplitude. (R mult = .9345, P <.001). Multiple regression equations were developed for predicting Factors I and II on the basis of the relevant physical voice characteristic measures.
dc.format.extentvii, 86, [10] leaves, bound : illustrationsen_US
dc.format.mediumFormat: Printen_US
dc.language.isoengen_US
dc.relation.ispartofTexas Christian University dissertationen_US
dc.relation.ispartofAS38.H63en_US
dc.subject.lcshRecognition (Psychology)en_US
dc.subject.lcshVoiceen_US
dc.titlePhysical and psychological correlates of speaker recognitionen_US
dc.typeTexten_US
etd.degree.departmentDepartment of Psychology
etd.degree.levelDoctoral
local.collegeCollege of Science and Engineering
local.departmentPsychology
local.academicunitDepartment of Psychology
dc.type.genreDissertation
local.subjectareaPsychology
dc.identifier.callnumberMain Stacks: AS38 .H63 (Regular Loan)
dc.identifier.callnumberSpecial Collections: AS38 .H63 (Non-Circulating)
etd.degree.nameDoctor of Philosophy
etd.degree.grantorTexas Christian University


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record